7805 overheating -- spa controller

Discussion in 'The Projects Forum' started by gmunee, Jan 2, 2009.

  1. gmunee

    Thread Starter New Member

    Jan 2, 2009
    I was helping my dad revive his old spa during the holiday. After some minor plumbing work, the spa was biased on. The control board works fine for 30 seconds, then goes unstable. There is a 7805 being used to regulate 18V to 5V for chip bias that is overheating and dropping out. There are only a few items being run off the 5V bus, and 7805's have a 2A rating. (see list of items on 5V bus below)
    I am guessing that maybe the output cap has become leaky overtime?
    Its got a big 2200uf, 35V electrolytic for an input cap and a tant 104 on the output. THink its the electrolytic. I dont know how to check ESR.
    Is there a way to determine where the current is going?
    Hooking a current probe to the output leg of the 7805 would determine the total draw, I think?
    Any ideas. More details below.

    The board worked 12 months ago, and has just sit. I cant see buying a new one for $500 from a spa dealer for a board that has a $50 bom cost.

    I dont have a schematic, customer serive guys wouldnt give me one either, but studying the basic functionality it can described it as follow. All through hole parts, easy to probe :).
    AC in to transformer primary, secondaries run through a bridge to generate 18V.
    18V use for coil voltage on various relays (pump, heater, light, blower)
    7805 used to generate 5 from 18V. 5V used to bias various chips.
    Here is what is on the 5V bus: HD404316D19S(4 bit controller), two HCF4001BE (NOR chips), HEF4521BP(freq divider), ULN2003 (tranistor pack), and a little comparator chip.

    Thanks for any help.
  2. bertus


    Apr 5, 2008

    Tantal capacitors can create a shortcut afther some time. (they sometimes even blow apart).
    What happens when you replace the tantalium capacitor?
    This capacitor is probably placed at the output of the 7805 to protect it from oscillating.

  3. gmunee

    Thread Starter New Member

    Jan 2, 2009
    I remove the tant cap right near the regulator, likely the output cap, and bias on. I figure this would still be a valid test, as it would likely only increase the ripple. It still overheated and shutdown. There are several 104 tantilums near each chip, likely for bypass. SHould I swap all these out for some standard ceramics?

  4. mik3

    Senior Member

    Feb 4, 2008
    The LM7805 is rated for 1A and not for 2A. Attach a heat sink on the 7805 to dissipate heat faster.
  5. hgmjr

    Retired Moderator

    Jan 28, 2005
    It sounds like the 7805 is overheating and then shutting itself down when the internal overheating protection circuitry is triggered due to excess heat.

    One thing to keep in mind is that the difference in the input voltage (18V) and the output voltage (5V) is being dropped across the 7805. The 13 volt delta means that for every approximate 80 milliamps of load current, the device will be subjected to a Watt of heating.

    Does the 7805 have a heatsink attached to it? If so then it may not be operating at its peak efficiency if it is not thermally coupled to the heatsink using thermal grease.

    If you find that the replacement of the capacitors does not remedy the problem then you can take a look at the heatsink attachment and make sure that it is well coupled.

  6. eblc1388

    AAC Fanatic!

    Nov 28, 2008
    Monitor the 5V bus with a DC voltmeter and then applying power to the circuit.

    Does 5V begin to drop or change after 30 seconds or sooner?

    Place your finger on various chips in turn to detect if they become unbearably hot right after startup. Keep touching each one until you have found one that's too hot.
  7. SgtWookie


    Jul 17, 2007
    I'd suspect the tantalum caps.

    If tantalum caps sit unused for a long time, they can short unexpectedly when again subjected to normal operating voltages. They usually go out with a "bang", but if you had a number of them short circuit, they might carry the current necessary to overheat the 7805.
  8. gmunee

    Thread Starter New Member

    Jan 2, 2009
    There is no heatsink on the 7805. I'll add a heatsink, change out the tantilums for ceramics, then I post back.

    Thanks for the help thus far. This is an very active forum. I'll sneak into work tomorrow to use the desoldering station and the metcal to make the rework a bit more fun.

    Stay tuned.
  9. hgmjr

    Retired Moderator

    Jan 28, 2005
    You may need to get your hands on some heat-sink compound. It is pretty messy but very useful in getting the most out of your heatsink.

  10. SgtWookie


    Jul 17, 2007
    Radio Shack carries a couple of heatsinks, and also small tubes of heatsink compound. The tab of the 7805 regulator is connected to the GND terminal.

    Make certain that the heatsink and the tab of the 7805 regulator are clean before you apply the heatsink compound. 91% Isopropyl alcohol works well for this; you can get it at almost any drugstore. The 70% has too much water in it; seems like it takes forever to dry.

    It just takes a thin coat of the heatsink compound. Use a small piece of cardboard, matchbook cover, tounge depressor or something else that's disposable. Try not to get it on your hands or clothes; it can be a pain in the neck to get off.

    It's the mechanical connection that's most important. The heat sink compound just fills in the little gaps caused by the porosity of the metals.
  11. Søren

    Senior Member

    Sep 2, 2006

    ESR has nothing to do with it, unless it is waaayyy high.
    Are you absolutely certain that the 100 nF cap are tantalum? I would expect e.g. an MKT cap or similar.

    A very good reason to expect the initial design to be working without adding a heatsink, just find the faulty conponent(s), whether it is the regulator, a cap, the diode bridge or perhaps a faulty connection somewhere (salts can build up and moisture might be another likely sinner, given the application).

    You don't need it for something this simple, but do you have a DMM?

    Others have mentioned the voltage drop of 13V which is a bit high for any serious current drain, but the chips mentioned don't suck up too much juice - the controller is the most hungry of them.

    I really don't think they are tantalums and the chances that one of them should be the culprit are small.

    The main thing to keep in mind here IMO, is that it used to work, so forget radical changes - Find what has gone awry, replace and enjoy a dip with a glass of bubbles for an overall inside out (or the other way 'round) experience (spice up with loose women for added fun ;))